Fossil Polygonal Periglacial Structures in Flanders (Belgium)

Fossil polygonal periglacial structures in Flanders (Belgium)

I. Heyse & G. Ghysels Department of Geography, Physical Geography, Ghent University, Ghent, Belgium

ABSTRACT: A systematic survey of aerial photographs in Flanders reveals an ephemeral visible polygonal pattern on the dissected plateau between the Coastal Plain and the Flemish Valley. The paper describes the characteristics of this pattern and discusses the origin and the age of these phenomena. The study is based on a comparative study of the literature, the analysis of aerial photographs, coupled with intensive fieldwork based on excavations. The spatial distribution of the sites, the visibility and the spatial structure of the patterns are described. Based on air photos the borders of the polygonal patterns were reconstructed and trenches were excavated in order to find underlying field evidence. The pattern corresponds with the infillings of frost wedges and excavated structures show characteristics of composite and ice-wedge pseudomorphs. The structures suggest formation in a continuous permafrost environment as a result of thermal contraction cracking. These features have been constrained to the period between the Elsterian and Weichelian, but are most probably of Late Pleistocene age.

1 INTRODUCTION

1.1 Definition of the problem

In Flanders (Belgium), various aerial photographs suggest the presence of polygonal networks (Fig. 2a), such as on the plateau between the Coastal Plain and the Flemish Valley (Figs. 1a, b). In this paper the origin and the age of these phenomena are discussed. According to Washburn (1973), polygonal networks have a worldwide occurrence. The largest polygons however, are restricted to cold periglacial and warm Figure 1a. General localisation of the study area. semi-arid environments. All the polygonal networks visible in the present- day temperate regions of the mid-latitudes are consid- ered as imprints from differential crop ripening of an underlying network of fossil frost-wedges, developed under palaeo-periglacial conditions (Christensen 1974; Gozdzik 1994; Johnson 1990; Kozarski 1993, 1995; Morgan 1971, 1972; Romanovskij 1973; Svensson 1972, 1973, 1976, 1984, 1988, 1990; Vandenberghe 1992; Walters 1994; Washburn 1973). Until now no evidence for these phenomena has been found in northern Belgium despite their presence in other European countries such as Poland and the U.K. This detailed study of networks is based on the aerial photograph archive of Prof. Dr. Bourgeois (RUG, Department of Archaeology and Old History of Europe) that was gradually set up with the assis- tance of J. Semey. The term “frost-wedge” is used for all wedge structures that originate due to thermal contraction in a frozen soil (French 1996; Kolstrup & Mejdahl 1986). The distinction between a permanent frost-wedge, developed in permafrost and a seasonal frost-wedge (“soil-wedge”), developed in a seasonally frozen Figure 1b. Localisation of the sites where research was ground is critical (Romanovskij 1985). carried out.

395 According to the primary infilling, distinctions are The controlling factors influencing their visibility made between ice-wedges, filled with ice, a mineral- are the thickness of the surface Quaternary cover, the wedge, filled with mineral material, and a composite- soil moisture differences, the contrast in lithology wedge, that consists of a mixture of both ice and mineral between host and fill materials, the type of growth, the material. On the basis of the degree of deformation of fos- period of the year and the meteorological conditions sil frost-wedges, a distinction is made between a “cast”, before and during the survey of the aerial photographs. that bears some resemblance to the original form and a “pseudomorph” that bears little resemblance to the origi- 2.2 The spatial distribution nal form (French 1996, p. 244). All the sites where polygonal networks occur are situ- 1.2 Methodology ated on the plateau between the Coastal Plain and the Flemish Valley (Fig. 1b). They are mainly located on This paper presents an analysis of aerial photographs low-angled (1°), NW – NE orientated slopes, some- coupled with intensive fieldwork, based on excavations. times on the interfluves and the rims of valleys. Firstly, all the oblique aerial photographs were geo- referenced with the help of the software “ILWIS” 2.3 The spatial structure (RUG, Department of Regional Geography, Prof. Dr. Antrop M.). Subsequently, the polygonal networks Based on the size, the geometry and the spatial were mapped cartographically with the help of the pattern of the polygonal structures 6 types have been software “Arc View” (Fig. 2b). The aim of the geo- metrical transformation was to obtain correctly orientated and full-sized polygonal networks. Polygonal patterns were reconstructed in the field and trenches were excavated in order to find underlying structures. The observations of the polygonal networks in the field were impeded by soil formation, homogenisation and a high groundwater level during wintertime. Special attention is paid to visible macroscopic structures, because they are of prime importance to understand the mechanism of infilling. A relative age for the features is obtained, based on the geomorphological position, the lithological con- Figure 2a. Polygonal networks visible in arable land on an texts and the relation to other sedimentary structures aerial photograph, site 1 (coordinates: X 83.35 km; of the polygonal patterns. Y 197.25 km; Lambert-72), Aalter, Oost-Vlaanderen, Belgium (Photo: Semey J., 07-08-1990, slidenumber 53 1.3 Study area 026, Department of Archaeology and Old History of Europe, RUG). Arrow: orientation (North). The study area is situated in the northern part of the plateau between the Coastal Plain and the Flemish Valley (Fig. 1). The rolling landscape slopes to the northeast, with mean heights of 20 m in the north and 50 m in the south. A scarp and vale structure reflects successive denudation in the Tertiary clayley and sandy substratum, infilled and levelled to some degree by Quaternary deposits (Heyse 1998).

2 ANALYSIS OF THE AERIAL PHOTOGRAPHS

2.1 Detection of polygonal networks

The polygonal networks are periodically visible in cultivated land (Fig. 2a). The outline of the polygons is explained by variation in the ripening of crops and as a consequence the features are therefore known as Figure 2b. Cartographic representation of the networks “crop marks” (Svensson 1982). visible in Figure 2a.

396 Table 1. Typology of polygonal networks in Flanders, with characteristics and site properties.

Figure 3. Wide sandy wedge structure developed in a tertiary clayley glauconiferous sandy substratum, Trench B-B (see Figure 2b and arrow in Figure 2a for localisation), site 1 (coordinates: X 83.35 km; Y 197.25 km; Lambert- 72), Aalter, Oost-Vlaanderen, Belgium. Length of the spade: 1.2m. See Table 2, site 1 for characteristics.

distinguished (Table 1). All 6 types belong to the “large type” (1 m) following Washburn’s (1973) classification. Figure 4. Drawing of the sand-wedge structure visible in According to various authors the networks could be Figure 3. Note the notch-like margins and large host inclu- developed by two manners: sions in the fill. • Owing to thermal contraction under cold periglacial conditions during the Pleistocene (Karte 1987; Pissart 1987; Vandenberghe & Pissart 1993). • As a result of desiccation under warm semi-arid conditions during the Pre-Quaternary period, analogous to those described by Neal et al. (1968, in Murton et al. 2000) in current warm semi-arid regions.

3 FIELD EVIDENCES Figure 5. Excavation A-B, site 15 (coordinates: X The polygonal crop marks are associated with sand- 80.00 km, Y 199.00 km; Lambert-72), Maria-Aalter, wedge structures (Figs 3, 4 and 6) in the subsoil. In out- Oost-Vlaanderen, Belgium. Sandy involutions developed in crops sand-wedge structures are visible with mean the tertiary green sandy glauconiferous substratum. Localisa- spacings of 3 to 10 m (Fig. 5). The term “sand-wedge” tion of Figure 5 (rectangle). See site 15 on Figure 1b for is used in a descriptive way, to indicate a sandy sedi- localisation of the excavation site. mentary structure. All the characteristics of the structures are summarised in Table 2. The tops of the is variable, ranging from wedge-shaped to strongly structures are between 0.4 m and 0.75 m depth, the bases irregular, sack-shaped structures. between 1.2 m and 1.8 m. The high width-height ratio, All the structures are filled with a complex of sandy varying between 1.5 and 2.7 is particular. The shape material with scattered pebbles (Figs 3, 4 and 6).

397 Table 2. Characteristics of the sand-involutions.

Site Type structure Height (cm)Width (cm) W/H Spatial Litho Chrono

1 Wedge 80 140 1.8 P s/Gs IV/III 4 Sack-shaped - C 60/90 240/115 2.7/1.92 P s/Gw IV/III 6 Deformed sack 115/75 100 2 P s/u IV/III 8 Sack-shaped 65 125 1.9 P s/l IV/IV 15a Deformed sack - C 120 100 1.7 P* s/Gs Sst IV/III 15b Wide sack 100 90 0.9 P* s/Gs IV/III 15c Long-waved sack - C 115 600 5.2 P* s/Gs IV/III

Legend: C: composite structure, W/H: width-height ratio, P: polygonal structure, *: sites with polygonal patterns in the neigh- bourhood, not observed at the site itself. Lithological units: s: sand, w: clayey sand, u: clay, l: loam, G: glauconite, Sst: sandstone fragments. Chrono: III: Tertiary, IV: Quaternary. Remark: structure 15c is an obliquely cut structure.

fill with large host inclusions (Figs 3, 4 and 6), a concave downward curved lamination/inclusions (Fig. 6) and scattered pebbles (1–10 cm) (Figs. 3, 4 and 6). Some of the structures are composite in nature, with supplementary wedge structures developed in the sand involutions (Fig. 6), sometimes piercing into the underlying host.

4 DISCUSSION

4.1 Origin

The polygonal sand-wedge patterns have been interpreted as a polygonal network of frost-wedge pseudomorphs. This is based on their geomorphological position, the denudation of the landscape during the Quaternary, the palaeoenvironmental and palaeolitho- logical context, the spatial structure of the patterns and the characteristics of the associated sand-wedges. Due to the intense denudation of the landscape in Flanders during the Quaternary it is most unlikely that these are Pre-Quaternary desiccation structures. Moreover, the structures are clearly filled with Qua- Figure 6. U-shaped, irregular sandy involution (dashed line) ternary sands, of which the exact stratigraphical age is in the Tertiary green glauconiferous sandy substratum (see not yet established. Frost-related features including Figure 4 for localisation). At the bottom right a green convex curved sandtongue (B) is formed in the ochre sandy infill. cryoturbations, fossil ice-wedges have been described Striking are the concave/downward curved green sandy by Heyse (1979, 1983, 1998, 1999) in the surrounding inclusions (A), which point to slump-, subsidence- and flow area. The structures are indicative for cold periglacial related processes. In the central part a supplementary, conditions in the area during the Weichsel. wing-formed wedge structure (C), tapering down to a point is All the aforementioned features point to slump, flow, formed. Site 15 (coordinates: X 80.00 km, Y 199.00 km; and subsidence related processes associated with the Lambert-72), Maria-Aalter, Oost-Vlaanderen, Belgium). See melt-out of an ice-body or ice-rich sediment (Harry & Table 2, site 15a for characteristics. Gozdzik 1988, Kolstrup 1987). Although gravel ele- ments might be present in frost-wedges with a primary infilling (Murton et al. 2000) or a secondary seasonal They are formed in different kinds of sediments infilling (Romanovskij 1973), the presence of large (sand, sandy clay, clayey sand, clay). host inclusions (Figs 3, 4 and 6) and a downward curved Common features are a downward curved lamination lamination of the fill (Fig. 6) may exclude such a and micro-stratification of the host material, irregular seasonal origin. margins (notch-like margins (Figs 3 and 4), tongue-like The structures are analogous with those observed structures (Figs 3, 4 and 6), small-scale faulting and a by Murton & French (1993c) in partly formed

398 ice-wedge and composite-wedge pseudomorphs in 5 CONCLUSIONS actual periglacial areas. All the previously mentioned sedimentary structures seem to be characteristic of 1 Polygonal structures are clearly present in Flanders composite and ice-wedge pseudomorphs. Presumably, (Belgium) on the plateau between the Coastal Plain the sand-wedges observed in the field are composite- and the Flemish Valley, irrespective of the lithol- and ice-wedge pseudomorphs. ogy of the substratum and the geomorphological It’s likely that considerable thaw consolidation has position. taken place giving wide, shallow wedge structures. 2 All the polygonal networks were caused by thermal High ice contents of host materials would explain such contraction of frozen ground associated with irregular forms (Murton & French 1993c). The strong palaeo-periglacial conditions. deformation and irregular appearance of some struc- 3 The polygonal crop marks are associated with sand tures is likely to be the result of thaw transformation wedges in the subsoil. They are interpreted as (Murton & French 1993c), slope-related processes frost-wedge pseudomorphs (type composite wedge (Gozdzik 1994, Mackay 1990), frost-heave and loading and ice-wedge), which may have been deformed by processes (Van Vliet-Lanoë 1988). thaw-, slope- and frost-related processes. For a more conclusive interpretation, texture analysis, 4 The visibility on aerial photographs is periodical the study of the grain morphology, microstructural and is related to specific soil conditions. The late analysis and the establishment of a regional framework spring and summer were the most favourable periods are all necessary. for observation. 5 In a denudation relief it is difficult to date the structures, but in the context of established regional cli- 4.2 Age mate reconstructions, the structures are constrained to the Elster-Weichsel period, and are probably In an eroded landscape it is very difficult to determine Late Pleistocene in age. the age of the structures. An absolute age determination 6 The high frequency of composite and ice-wedge is almost impossible due to the absence of datable hori- pseudomorphs in northwest Flanders in various zons, such as peat layers and other organic remnants sediments and in different geomorphological posi- (Svensson 1988, p. 65). The application of luminescence tions points to the presence of continuous per- dating is potentially a very promising technique in mafrost during the Late Pleistocene, as suggested obtaining an (in)direct age of the fossil wedge struc- by palaeoclimate reconstructions (Vandenberghe tures however. Nevertheless, based on the geomor- et al. 1998). phological and lithostratigraphical position and the relation to other sedimentary structures it is possible to obtain a minimum–maximum age. ACKNOWLEDGEMENTS All the sites are located in the northern part of the interfluve between the Coastal Plain and the Flemish We wish to thank all the following people and Valley, of which the highest parts correspond with departments: interfluvial terraces formed during the Elster Glacial Professor Dr. Bourgeois J. and Semey J. (RUG, (Oxygen Isotope Stage 12), resting upon the Tertiary Department of Archaeology and Old History of Europe) substratum of Eocene age. for the disposal of the aerial photographs. The Consequently, the structures were developed during Department of Regional Geography (Professor the Elster Glacial (Oxygen Isotope Stage 12), or after- Dr. Antrop M., RUG) for use of the software ILWIS and wards, during one of the later glacial periods. In gen- Arc View. Professor C. Harris & Professor W. Haeberli eral one can assume that they were formed during a for their useful comments on the initial manuscript. Late-Pleniglacial period (Oxygen Isotope Stage 2). According to Vandenberghe & Pissart (1993) a Late- Glacial age should not be ruled out. Permafrost condi- REFERENCES tions necessary for the formation of ice-wedges and composite-wedges existed during different parts of Christensen, L. 1974. Crop-marks revealing large-scale the Late Pleistocene (Vandenberghe J. et al. 1998, patterned ground structures in cultivated areas, south- Vandenberghe & Pissart 1993). western Jutland, Denmark. Boreas 3: 153–180. French, H.M. 1996. The periglacial environment. London: Numerous structures are composite in nature and Longman. point to a polygenetic and polycyclic character. The Gozdzik, J.S. 1986. Structures de fentes à remplissage composite structures were presumably developed due primaire sableux du Vistulien en Pologne et leur to reactivation during different glacial periods and/or importance paleogéographique. Biuletyn Peryglacjalny stadials. 31: 71–105.

399 Gozdzik, J.S. 1994. Etudes des fentes de gel en Pologne Neal, J.T., Langer, A.M. & Kerr, P.F. 1968. Giant dessica- centrale. Biuletyn Peryglacjalny 33: 49–78. tion polygons of Great Basin playas, Geological Society Harry, D.G. & Gozdzik, J.S. 1988. Ice-wedges: growth, of America Bulletin 79: 69–90. thaw transformation, and paleoenvironmental signifi- Pissart, A. 1987. Weichselian periglacial structures and their cance. Journal of Quaternary Science 3: 39–55. environmental significance: Belgium, the Netherlands Heyse, I. 1979. Bijdrage tot de geomorfologische kennis and northern France. In: J. Boardman (ed), Periglacial van het noordwesten van Oost-Vlaanderen (België). processes and landforms in Britain and Ireland: Verhandelingen van de Koninklijke Academie voor 77–85. Cambridge: Cambridge University Press. Wetenschappen, Letteren en Schone Kunsten van België. Romanovskij, N.N. 1973. Regularities in formation of frost- Klasse der Wetenschappen. 155. fissures and development of frost-fissure polygons. Heyse, I. 1983. Cryoturbation types in eolian Würm Late Biuletyn Peryglacjalny 23: 237–277. Glacial sediments in Flanders, Belgium. Polarforshung Romanovskij, N.N. 1985. Distribution of recently active ice 53: 87–95. and soil wedges in the U.S.S.R. In: M. Church & Heyse, I. 1998. An environmental model of valley damming O. Slaymaker (eds), Field and Theory, Lectures in due to eolian action during the Late Weichselian glacial Geocryology.: 154–165. Vancouver: University of British period. Data based on a pipe-line section study (Ede Columbia Press. Valley – Belgium). Natuurwetenschappelijk Tijdschrift Svensson, H. 1972. The use of stress situations in the vege- 78: 149–159. tation for detecting ground conditions on aerial photo- Heyse, I. 1999. Fossil periglacial remnants in the graphs. Photogrammetria 28: 75–87. Beernem-Mouton excavation in Flanders (Belgium). Svensson, H. 1973. Distribution and chronology of relict Biuletyn Peryglacjalny 38: 53–68. polygon patterns on the Laholm Plain, the Swedish Johnson, W.H. 1990. Ice-wedge casts and relict patterned west coast. Geografiska Annaler 54A: 159–175. ground in central Illinois and their environmental Svensson, H. 1976. Relict ice-wedge polygons revealed on significance. Quaternary Research 33: 51–72. aerial photographs from Kaltenkirchen, northern Karte, J. 1987. Pleistocene periglacial conditions and geo- Germany. Geografisk Tidsskrift 75: 8–12. morphology in north central Europe. In: J. Boardman Svensson, H. 1982. The use of aerial photographs and remote (ed), Periglacial processes and landforms in Britain sensing techniques in research on fossil periglacial and Ireland: 67–75. Cambridge: Cambridge University features. Biuletyn Peryglacjalny 29: 129–138. Press. Svensson, H. 1984. The periglacial form group of south- Kolstrup, E. & Mejdahl, V. 1986. Three frost wedge casts western Denmark. Geografisk Tidsskrift 84: 25–34. from Jutland (Denmark) and TL dating of their infill. Svensson, H. 1988. Ice-wedge casts and relict polygonal Boreas 15: 311–321. patterns in Scandinavia. Journal Quaternary Science Kolstrup, E. 1987. Frost wedge casts in western Jutland and 3: 57–67. their possible implications for European periglacial Svensson, H. 1990. Relict periglacial structures. Occurrences, research. Zeitschrift für Geomorphologie 31: 449–461. age and development in different matrices on a coastal Kozarski, S. 1993. Late Plenivistulian deglaciation and the plain of southwestern Sweden. Geografiska Annaler expansion of the periglacial zone in NW Poland. 72A: 79–91. Geologie en Mijnbouw 72: 143–157. Vandenberghe, J. 1992. Cryoturbations: a sediment struc- Kozarski, S. 1995. The periglacial impact on the deglacia- tural analysis. Permafrost and Periglacial Processes tion area of northern Poland after 20 kyr BP. Biuletyn 3: 343–352. Peryglacjalny 34: 73–102. Vandenberghe, J. & Pissart, A. 1993. Permafrost changes in Mackay, J.R. 1990. Some observations on the growth Europe during the Last Glacial. Permafrost and and deformation of epigenetic, syngenetic and anti- Periglacial Processes 4: 121–135. syngenetic ice wedges. Permafrost and Periglacial Vandenberghe, J., Kasse, K. & Coope, R. (eds) 1998. Processes 1: 15–30. Climatic reconstruction of the Last Interglacial- Morgan, A.V. 1971. Polygonal patterned ground of the Glacial Cycle in Western & Central Europe (Special Late Weichselian age in the area north and west of Issue of the Journal of Quaternary Science 13 (issue Wolverhampton, England. Geografiska Annaler 53A: 5): 361–497. Chichester: John Wiley & Sons. 146–156. Van Vliet-Lanoë, B. 1988. The significance of cryoturbation Morgan, A.V. 1972. Late Wisconsinan ice-wedge polygons phenomena in environmental reconstruction. Journal near Kitchener, Ontario, Canada, Canadian Journal of Quaternary Science 3: 85–96. of Earth Sciences 9: 607–617. Walters, J.C. 1994. Ice-wedge casts and relict polygonal Murton, J.B. & French, H.M. 1993c. Thaw modification of patterned ground in north-east Iowa, USA. Permafrost frost-fissure wedges, Richards island, Pleistocene and Periglacial Processes 5: 269–282. Mackenzie Delta, western Arctic Canada. Journal Washburn, A.L. 1973. Periglacial processes and environ- of Quaternary Science 8: 185–196. ments. London: Edward Arnold. Murton, J.B., Worsley, P. & Gozdzik, J. 2000. Sand veins and wedges in cold Aeolian environments. Quaternary Science Reviews 19: 899–922.

400